4 ms·

Read the novel "The Right Stuff" for more background. There was a disagreement as to whether astronauts would be pilots of their spacecraft or mostly along for

by dockd 11y ago

Read the novel "The Right Stuff" for more background. There was a disagreement as to whether astronauts would be pilots of their spacecraft or mostly along for the ride. I wonder if any of that desire was responsible for the Space Shuttle.

troutwine 11y ago
Also, Digital Apollo by David Mindell. Early on--being the late 50s and into the early 70s--the desire for pilot control was entirely the rational thing. Automation was very primitive and pretty much limited to doing the right thing only when the situation was nominal. Pilots are easily overwhelmed by tedium and a "pilot in the loop" design methodology prevailed at NASA. Basically that means if a human can do it, they should. The Soviet design took a different tack and invested more in automation, dooming some cosmonauts but being probably for the best long-term. Nowadays we know how to do some things very, very well--rendezvous--so heavy automation makes sense. In other areas that are more research projects pilot in the loop still makes sense. It's a tradeoff of efficiency for flexibility and fault tolerance.
- snowwrestler 11y ago
  It looks to me, as an amateur observer, that the automation vs. pilot-in-the-loop debate is today not yet settled. The Air France crash into the Atlantic was entirely caused by pilots who were used to automation, and made a series of wrong decisions when it cut out on them, even though it cut out according to spec (i.e. it did not fail, just handed control to humans as designed). It's arguable that had human pilots been actively flying that plane the whole time, it would not have crashed.
  - troutwine 11y ago
    Absolutely. It'll never be settled, so long as human ambition pushes us into unknowns. You can automate what you understand perfectly and make trade-offs in areas where you don't, as above. Fascinating subject, really.

by dockd 11y ago

troutwine 11y ago

Also, Digital Apollo by David Mindell. Early on--being the late 50s and into the early 70s--the desire for pilot control was entirely the rational thing. Automation was very primitive and pretty much limited to doing the right thing only when the situation was nominal. Pilots are easily overwhelmed by tedium and a "pilot in the loop" design methodology prevailed at NASA. Basically that means if a human can do it, they should. The Soviet design took a different tack and invested more in automation, dooming some cosmonauts but being probably for the best long-term. Nowadays we know how to do some things very, very well--rendezvous--so heavy automation makes sense. In other areas that are more research projects pilot in the loop still makes sense. It's a tradeoff of efficiency for flexibility and fault tolerance.

snowwrestler 11y ago
It looks to me, as an amateur observer, that the automation vs. pilot-in-the-loop debate is today not yet settled. The Air France crash into the Atlantic was entirely caused by pilots who were used to automation, and made a series of wrong decisions when it cut out on them, even though it cut out according to spec (i.e. it did not fail, just handed control to humans as designed). It's arguable that had human pilots been actively flying that plane the whole time, it would not have crashed.
- troutwine 11y ago
  Absolutely. It'll never be settled, so long as human ambition pushes us into unknowns. You can automate what you understand perfectly and make trade-offs in areas where you don't, as above. Fascinating subject, really.